Example usage for org.apache.commons.math.distribution GammaDistributionImpl GammaDistributionImpl

Introduction

In this page you can find the example usage for org.apache.commons.math.distribution GammaDistributionImpl GammaDistributionImpl.

Prototype

public GammaDistributionImpl(double alpha, double beta) 

Source Link

Document

Create a new gamma distribution with the given alpha and beta values.

Usage

From source file:org.renjin.stats.internals.Distributions.java

@DataParallel
@Internal/*from  ww w .  java  2s . c  o  m*/
public static double pgamma(@Recycle double q, @Recycle double shape, @Recycle double scale, boolean lowerTail,
        boolean logP) {
    return p(new GammaDistributionImpl(shape, scale), q, lowerTail, logP);
}

---

From source file:org.renjin.stats.internals.Distributions.java

@DataParallel
@Internal// ww w. jav a 2 s  .  c o  m
public static double qgamma(@Recycle double p, @Recycle double shape, @Recycle double scale, boolean lowerTail,
        boolean logP) {
    return q(new GammaDistributionImpl(shape, scale), p, lowerTail, logP);
}

---

From source file:timsaddons.modelselection.ModelSwitchOperator.java

@Override
public double proposal() {

    // Uniformly select new model from available models:
    int oldModel = modelNumberInput.get().getValue();
    int newModel;
    do {//from ww  w .ja v a 2s  . c  o m
        newModel = Randomizer.nextInt(nModels);
    } while (newModel == oldModel);

    // Update model number
    modelNumberInput.get().setValue(newModel);

    // Record selection probability of old parameters in HR:
    double logHR = 0;

    for (int i = 0; i < parameterModelIndicesInput.get().getDimension(); i++) {
        int m = parameterModelIndicesInput.get().getValue(i);

        if (m == oldModel) {
            double shape = proposalShapesInput.get().getValue(i);
            double scale = proposalMeansInput.get().getValue(i) / shape;
            double x = parametersInput.get().get(i).getValue();

            logHR += (new GammaDistributionImpl(shape, scale)).density(x);
        }

        if (m == newModel) {
            double shape = proposalShapesInput.get().getValue(i);
            double scale = proposalMeansInput.get().getValue(i) / shape;
            double xprime = Randomizer.nextGamma(shape, 1.0 / scale);

            logHR -= (new GammaDistributionImpl(shape, scale)).density(xprime);

            parametersInput.get().get(i).setValue(xprime);
        }
    }

    return logHR;
}

---

From source file:timsaddons.modelselection.ModelSwitchOperator.java

/**
 * Main for debugging.//w w  w.  j  a  v a 2s. co  m
 * 
 * @param args 
 */
public static void main(String[] args) throws FileNotFoundException {

    // Check that RNG is working as expected
    PrintStream outf = new PrintStream("rngtest.txt");
    for (int i = 0; i < 10000; i++) {
        outf.println(Randomizer.nextGamma(2.0, 3.0));
    }
    outf.close();

    // Check that Gamma distribution PDF is behaving as expected:
    outf = new PrintStream("pdftest.txt");
    outf.println("x p");
    for (double x = 0; x < 10; x += 0.01) {
        outf.println(x + " " + (new GammaDistributionImpl(2.0, 1.0 / 3.0)).density(x));
    }
    outf.close();
}

---

From source file:yabby.evolution.sitemodel.SiteModel.java

/**
 * discretisation of gamma distribution with equal proportions in each
 * category/*w  ww .j a  v a2  s. c  om*/
 * @param node
 */
protected void calculateCategoryRates(final Node node) {
    double propVariable = 1.0;
    int cat = 0;

    if (/*invarParameter != null && */invarParameter.getValue() > 0) {
        if (hasPropInvariantCategory) {
            categoryRates[0] = 0.0;
            categoryProportions[0] = invarParameter.getValue();
        }
        propVariable = 1.0 - invarParameter.getValue();
        if (hasPropInvariantCategory) {
            cat = 1;
        }
    }

    if (shapeParameter != null) {

        final double a = shapeParameter.getValue();
        double mean = 0.0;
        final int gammaCatCount = categoryCount - cat;

        final GammaDistribution g = new GammaDistributionImpl(a, 1.0 / a);
        for (int i = 0; i < gammaCatCount; i++) {
            try {
                // RRB: alternative implementation that seems equally good in
                // the first 5 significant digits, but uses a standard distribution object
                if (useBeast1StyleGamma) {
                    categoryRates[i + cat] = GammaDistributionQuantile((2.0 * i + 1.0) / (2.0 * gammaCatCount),
                            a, 1.0 / a);
                } else {
                    categoryRates[i + cat] = g
                            .inverseCumulativeProbability((2.0 * i + 1.0) / (2.0 * gammaCatCount));
                }

            } catch (Exception e) {
                e.printStackTrace();
                System.err.println("Something went wrong with the gamma distribution calculation");
                System.exit(-1);
            }
            mean += categoryRates[i + cat];

            categoryProportions[i + cat] = propVariable / gammaCatCount;
        }

        mean = (propVariable * mean) / gammaCatCount;

        for (int i = 0; i < gammaCatCount; i++) {

            categoryRates[i + cat] /= mean;
        }
    } else {
        categoryRates[cat] = 1.0 / propVariable;
        categoryProportions[cat] = propVariable;
    }

    ratesKnown = true;
}

---

• «
• 1
• 2
• 3
• »